International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
DOI : 10.5121/ijwest.2011.2402 21
Semantic – Based Querying Using Ontology in
Relational Database of Library Management
System
Ayesha Banu1
, Syeda Sameen Fatima2
and Khaleel Ur Rahman Khan3
1
Asso.Professor, Dept. of Informatics, Alluri Institute of Management Sciences,
KU, Warangal.
ayeshaaims@gmail.com
2
Professor, Dept. of C.S.E , Univ. College of Engineering ,OU , Hyderabad
sameenf@gmail.com
3
Professor, Dept.of C.S.E, ACE Engineering College , JNTU, Hyderabad
khaleelrkhan@aceec.ac.in
ABSTRACT:
The traditional Web stores huge amount of data in the form of Relational Databases (RDB) as it is good at
storing objects and relationships between them. Relational Databases are dynamic in nature which allows
bringing tables together helping user to search for related material across multiple tables. RDB are
scalable to expand as the data grows. The RDB uses a Structured Query Language called SQL to access
the databases for several data retrieval purposes. As the world is moving today from the Syntactic form to
Semantic form and the Web is also taking its new form of Semantic Web. The Structured Query of the RDB
on web can be a Semantic Query on Semantic Web. The SPARQL is the Query Language recommended by
W3C for the RDF(Resource Description Framework). RDF is a directed, labeled graph data format for
representing information in the Web and is a very important layer of the Semantic Web Architecture. In this
paper we consider the Library Management System (LMS) database, taking some tuples of the LMS
Relational Schema. We discuss how the RDF code is scripted and validated using RDF Validator and how
RDF Triples are generated. Later we give the graphical representation of the RDF triples and see the
process of extracting ontology from the RDF Schema and application of the Semantic Query.
KEYWORDS:
Semantic Web, Ontology, Semantic Query, Relational Database.
1. INTRODUCTION
Today’s Web is a big pool of information stored in various forms. The World Wide Web has also
changed the way people communicate with each other. Most of the web content today is suitable
for human consumption with typical uses involving searching for the required information,
reviewing the online stores and placing order for products and many more. The main tools people

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
22
use to search today’s web are the keyword based search engines like Google and Yahoo. Even if
the search is successful the person has to still browse the retrieved documents further to extract
the exact information he is looking for. The main obstacle to provide a better support to web user
is that “the meaning of the web document is not machine-accessible”.
There are tools which can retrieve texts , count number of words and check their spelling but
when it comes to interpreting the sentences and extracting information useful to users the
capability of current software’s is limited. The solution for this is provided by the Semantic Web
where traditional Web is focused on people, the Semantic Web is focused on machines.
The Web requires a human operator, using computer systems to perform the tasks required to
find, search and aggregate its information. It's impossible for a computer to do these tasks without
human guidance because Web pages are specifically designed for human readers. The Semantic
Web aims to change it by presenting Web page data in such a way that it is understood by
computers, enabling machines to do the searching, aggregating and combining of the Web’s
information — without a human operator. The organization of this paper is as follows: Section 2
presents the concept of Semantic Web and its major layers. Section 3 focuses on the Related
Work for the process of building ontology based query for the relational database (Library
Management System Database). Section 4 describes the Semantic Query in the RDB and Section
5 concludes the paper.
2.SEMANTIC WEB CONCEPTS AND BACKGROUND
In this section we describe the basic concepts of the Semantic Web, its layered architecture,
different languages for the semantic web , ontology , RDF Triple RDS(Relational Database
Scheme) and RDFS(RDF Schema).
2.1. The Semantic Web
The Semantic Web (SW) is an extension of the current web where the information is presented in
a well-defined manner, better enabling computers and people to work in cooperation.[1] Berners-
Lee suggested a layer structure for the Semantic Web as shown below:
[ Figure Source:en.wikipedia.org/wiki/Semantic_Web_Stack]

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
23
Figure 1: The Layered Architecture of Semantic Web
The architecture of the Semantic Web can be classified as four major parts
1. XML : The Representation Layer
2. RDF : The Knowledge Representation Layer
3. RDF Schema : The Ontology Layer
4. Agents : The Logic , Proof and Trust Layers
2.2. Languages for the Semantic Web
There are many languages proposed for semantic web where every language has its own utility
and style of scripting and working. Some of the languages for Semantic Web can be [2]:
HTML (Hyper Text Markup Language): is the standard language in which Web pages are
written. HTML was designed only for the purpose of displaying the data and it has no separation
between the logical structure of the web page and the actual data of the page. This raised the
difficulty in understanding the syntax and semantics.
XML (Extensible mark-up language): was designed as a language for mark-up or annotation of
documents. XML separates the logical structure of the data from data itself. An XML object is a
labeled tree and consists of objects with attributes and values. XML allows the definition of any
kind of annotation, which opens the way to annotation with ontologies. XML Schema allows the
definition of grammars for valid XML documents, but XML cannot recognize the semantics of
information.

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
24
RDF (Resource Description Framework): To recognize the semantics of information RDF is
proposed which represent information in a concept representation language. RDF is the widely
accepted standard proposed by W3C Consortium for representing metadata. RDF documents
consist of three types of entities: resources, properties, and statements. Resources may be Web
pages or parts of Web pages. Properties are specific attributes, characteristics, or relations
describing resources. Statements can the object–attribute–value triples.
RDF Schema (RDFS): defines a simple modeling language on top of RDF which includes
classes, is-a relationship between classes and between properties, and domain/range restrictions
for properties. RDF and RDF Schema are written in XML syntax.
OWL (Web Ontology Language): Like RDF and RDF Schema, OWL is a W3C recommendation,
intended to support more elaborate semantics. OWL includes elements from description logics
and provides many constructs for the specification of semantics, including conjunction and
disjunction, existentially and universally quantified variables and property inversion.
SPARQL: SPARQL (pronounced as "sparkle") is an RDF query language, stands for SPARQL
Protocol and RDF Query Language. It was standardized by the RDF Data Access Working
Group (DAWG) of the World Wide Web Consortium.
2.3. The Ontology
Ontology stands as the most important concept in Semantic Web, which can be defined as a
collection of key concepts and their inter relationship providing an abstract view of an application
domain. The Ontology enables both user and system to communicate with each other by the
shared and common understanding of a domain [3]. For the web, ontology is about the exact
description of web information or web resources and the relationship between them.
RDF (Resource Description Framework) can be used for describing the web resources and the
relationship between them. The fundamental concepts of RDF are Resources, Properties and
Statements.
Resources can be any object about which we would think. A resource may be an author or a book
or a publisher and so on. Every resource has a Uniform Resource Locator (URL) or Web
Address.
Properties are special kinds of resources which describe the relation between the resources. For
example “title” (title of a book), “age” (age of any person) and so on. These properties in RDF are
also identified by URLs.
Statements assert or declare the properties of resources. A statement in RDF is Object-Property-
Value triple consisting of a resource (Object), property (the property of the resource) and value (is
the property value of the resource) [4].
The graphical representation of a RDF triple can be shown as below:

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
25
Figure 2: Graphical Representation of RDF Triple
3. RELATED WORK
To understand how the RDF triples are formed , how the RDF code is produced and
validated and how the Graphical Representation is done, we consider the following
example table of 2 books with its title , author , and price and published year given in the
table.
Table 1: Book Details
Book Title Author Price PubYear
Data Mining Arun.K.Pujari $56 1999
Semantic Web Tim Berner’s Lee $65 2004
For the above data regarding books details the RDF code can be written as follows [8]:
<?xml version="1.0"?>
<rdf: RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:lib="http://www.libraryexample.cm/lib#">
<rdf:Description
rdf:about="http://www.libraryexample.cm/lib/DataMining">
<lib:author>Arun.k.Pujari</lib:author>
<lib:price>$56</lib:price>
<lib:pubyear>1990</lib:pubyear>
</rdf:Description>
<rdf:Description
rdf:about="http://www.libraryexample.cm/lib/SemanticWeb">
<lib:author>Tim Brener's Lee</lib:author>
<lib:price>$65</lib:price>
<lib:pubyear>2004</lib:pubyear>
</rdf:Description>
</rdf:RDF>
The above code can now be subjected against the RDF Validator and Converter to check for the
syntax validation of the code and obtain the RDF Triples. Demo version of the RDF validator can
Object Value
Property

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
26
be available at http://www.rdfabout.com/demo/validator/ .The results of RDF Validator and
Converter for the above code can be
RDF Validation Result
Syntax validated OK.
The document as Notation 3
@prefix lib: <http://www.libraryexample.cm/lib#>.
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.
<http://www.libraryexample.cm/lib/DataMining> lib:author "Arun.k.Pujari"; lib:price "$56";
lib:pubyear "1990".
<http://www.libraryexample.cm/lib/SemanticWeb> lib:author "Tim Brener's Lee"; lib:price "$65";
lib:pubyear "2004".
The underlying triples
<http://www.libraryexample.cm/lib/DataMining> <http://www.libraryexample.cm/lib#author>
"Arun.k.Pujari".
<http://www.libraryexample.cm/lib/DataMining> <http://www.libraryexample.cm/lib#price> "$56".
<http://www.libraryexample.cm/lib/DataMining> <http://www.libraryexample.cm/lib#pubyear> "1990".
<http://www.libraryexample.cm/lib/SemanticWeb> <http://www.libraryexample.cm/lib#author> "Tim
Brener's Lee".
<http://www.libraryexample.cm/lib/SemanticWeb> <http://www.libraryexample.cm/lib#price> "$65".
<http://www.libraryexample.cm/lib/SemanticWeb> <http://www.libraryexample.cm/lib#pubyear> "2004".
The Graphical representation of the RDF triples can result in a RDF Graph [9]. A part of RDF
Graph for above example can be as shown
Figure 3: Graphical Representation of our Example RDF Triple
3.1. RDB and RDFS
Relational Databases are used more to store the content of the website. The integration of XML
with relational database systems to enable the storage, retrieval, and update of data is of major

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
27
importance [5]. But today many of the Web-based information systems do not aim at purely
providing read-only access to their content which is simply represented in terms of web pages
stored in the web server’s directory but provide more access to their data. [6]. So Compared to
XML and relational databases, which are structural-oriented, RDF takes into consideration a
knowledge oriented approach that is designed specifically for the Web and that is extremely
useful for the Semantic Web. One of the advantages of RDF over XML and relational model is
that an RDF graph depicts in a unique form the information to be conveyed. Following table
shows the relationship between the Entity Relationship Model (Structural model) and the
ontology model (Semantic model) [3]
Table 2: The Relationship between ER-Model and Ontology Model
Graphical
Representation
Implementation
Schema
Instances Modeling
ER Model ER-Model Relational Schema Relational
Database
Data Modeling
Ontology Model Ontology RDF Schema RDF Knowledge
Modeling
4. BUILDING ONTOLOGY BASED QUERY
To build the ontology based query for the Library Management System database we consider the
framework proposed in[3] consisting of 2 phases, offline ontology extraction and online query
issuing. In offline ontology extraction, the system extracts the explicit classes and relations from
the relational schema. Then the domain expert will adapt the extracted ontology by adding the
implicit relations to complete the ontology. In online query operation the user can issue a
semantic query to the system, and the system maps that query into a related SQL query for the
underlining relational database. The framework can be as shown:
Figure 4: Framework for querying relational database based on ontology

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
28
4.1. Extracting Ontology from Relational Database
BOOK, MEMBER and PUBLISHER are the basic entities for the example database we use .
Every Library has books with unique id, title, author, price and status (available or not). Members
of the library borrow books and every member has member id, name, address and member type
(student/professor/researcher). Library also maintains the details of publishers of the books
including the publisher id, name and address. The database also maintains the complete book
details and issue details. The complete Relational Database of the Library Management System
considered can be shown in table below.
Table 3: A Relational Database for the Library Management System
Relation Primary Key Foreign Key
Book(Bid,Title,Author,Price,Status,Mid,Pid) Bid Mid ref. to Member
Pid ref. to Publisher
Member(Mid,Mname,Addr,Mtype) Mid
Publisher(Pid,Pname,Addr) Pid
IssueDetails(Mid,Mname,Bid,Title,Doi,Dor) Bid , Mid Bid ref. to Book
Mid ref. to Member
BooksDetails(Bid,Title,Price,Pid,Pname) Bid,Pid Bid ref. to Book
Pid ref. to Publisher
In the semi-automatic process of extracting ontology from relational database we apply some
rules like.
Rule1: If the primary key of any relation is unique and do not contain the primary key of any
other relation then we consider such relation as on ontological class.
As per our Library example Member and Publisher are separate Ontological classes.
lib:Member rdf:type rdfs:class.
(Mid,Mname,Addr,Mtype)
lib:Publisher rdf:type rdfs:class.
(Pid,Pname,Addr)
Rule2: If the foreign key of any relation R1 is the Primary key of any other relation R2 then
there exists an object property from R1 to R2 and the domain is R1 and range is R2.
As per our Library example members borrow books from library so Borrowerid from Borrow can
be an object property with domain as Member and range as Book.
lib:Borrow rdf:type rdf:property.
lib:Borrow rdfs:domain rdf:Member.
lib:Borrow rdf:range rdf:Book.
Many more rules of such type can be applied to extract the ontology from the relational database.
Following figure shows the extracted ontology where all the class and the data type properties are
defined but some object type properties are not yet defined.

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
29
Figure 5: The Extracted Wrapper Ontology
The next step will be to adapt the extracted ontology to pre-defined domain ontology. This stage
will add the explicit definition of the implicit relationships and adjust directions of the object
properties between classes. For example the property (relation) between the Book and Publisher
classes is not fully defined yet. Based on the domain ontology, the obtained full wrapper ontology
is shown as:
Figure 6: The Full Wrapper Ontology

International Journal of Web & Semantic Technology (IJWesT) Vol.2, No.4, October 2011
30
The following RDF Schema Triples notation is the part of the whole schema.
……….
……….
lib:Book rdf:type rdfs:class
lib:Member rdf:type rdfs:class
lib:Publisher rdf:type rdfs:class
lib:PublishedBy rdf:type rdf:property
lib:PublishedBy rdfs:domain rdf:Book
lib:PublishedBy rdf:range rdf:Publisher
……….
(Here lib is the namespace for the extracted ontology)
4.2. SEMANTIC QUERY IN RDB (Relational Data Base)
After extracting and refining the wrapper ontology, the user can issue semantic queries based on
extracted ontology concepts (Keywords) and these queries will be mapped onto plain syntactic
SQL queries. The semantic queries will be based on SPARQL(pronounced as "sparkle") is an
RDF query language, stands for SPARQL Protocol And RDF Query Language. SPARQL can be
used to express queries across diverse data sources, whether the data is stored natively as RDF or
viewed as RDF via middleware[7]. SPARQL contains capabilities for querying required and
optional graph patterns along with their conjunctions and disjunctions. SPARQL also supports
extensible value testing and constraining queries by source RDF graph. The results of SPARQL
queries can be result sets or RDF graphs.
Following is the SPARQL syntax query to find the member details (Mname, Mtype, and Addr)
who have borrowed the book titled “Data Mining” from the Library.
Prefix lib: http://www.libraryexample.cm
Select ?Mname ?Mtype ?Addr ?Title
Where {
?s lib:Mid ?p.
?p lib:Mname ?Mname.
?p lib:Mtype ?Mtype.
?p lib:Addr ?Addr.
?p lib:Borrow ?q.
?q lib:Bid ?r.
?r lib:Title?Title.
} ?Title=”Data Mining”
The above SPARQL query is translated to SQL query by following procedure [3].
The triples that share the same subject are grouped as they represent the same table
information. So, each group represents some information about one concept in mediated
ontology. For example, ?p subject group represent the Member information.
Based on the mapping information, the translation algorithm replaces all predicates in the
triples with corresponding columns name in relational databases tables. For example Mname,
Mtype, and Addr are there in the Member table.

International Journal of Web Semantic Technology (IJWesT) Vol.2, No.4, October 2011
31
If the predicate is not in the columns name, then it will be in object property names which
are related to the linking tables. For example the Borrow property is not their in any table column.
By inspection, it is in the mediated ontology linking the Member concept to the Book concept.
For each separate group, a subquery clause is created, which consists of three parts:
Select clause, From clause and Where clause. The Select clause is created according variable that
occurs both in triple and in SPARQL select clause. The From clause is created according the
column name in the triples. And the Where clause is created according the columns and mapping
information. After all clauses are created, they are combined to construct the complete a query
clause.
For the SPARQL query given above ,the SQL query translated is below.
Select Member.Mname , Member.Mtype , Member.Addr , Book.Title
From Member,Book
Where Member.Mid=Book.Mid and Book.Title=”Data Mining”;.
5. CONCLUSION
As we know that the traditional web includes a huge amount of Relational Databases (RDB) that
support structuring data on a syntactic base. Converting available data stored in relational
database into RDF format is tedious and error prone. So, instead of migrating available legacy
data in relational database into RDF format based on ontology, we can build a wrapper that act as
a translator from the semantic queries issued to the system into the syntactic data available within
these databases. So this approach of extracting ontology from the RDB and allowing user to issue
a semantic query and translating it into syntactic sql query and then transforming the results set
obtained from relational databases into RDF triples using namespaces is more beneficial. There
are many more concepts involved in mapping the semantic query using schema ontology into sql
statements which can also be considered for implementation.
6. REFERENCES
[1] Gerd Stumme, Andreas Hotho, Bettina Berendt , “Semantic Web Mining State of the Art and Future
Directions: A Survey” Web Semantics: Science, Services and Agents on the World Wide Web
Volume 4, Issue 2, June 2006, Pages 124-14
[2] Bettina Berendt, Andreas Hotho, Dunja Mladenic,Maarten van Someren, Myra Spiliopoulou, Gerd
Stumme , ” A Roadmap for Web Mining: From Web to Semantic Web”
[3] Mostafa E. Saleh , “Semantic-Based Query in Relational Database Using Ontology” , Canadian
Journal on Data, Information and Knowledge Engineering Vol. 2, No. 1, January 2011.
[4] Grigoris Antoniou, Frank Van Harmelen , “A Semantic Web Premier” , Second Edition, MIT Press.
[5] Gerti Kappel, Elisabeth Kapsammer and Werner Retschitzegger, Integrating XML and Relational
Database Systems, World Wide Web: Internet and Web Information Systems, 7, 343–384, 2004,
Kluwer Academic Publishers.

International Journal of Web Semantic Technology (IJWesT) Vol.2, No.4, October 2011
32
[6] D. Florescu, A. Levy, and A. Mendelzon, “Database techniques for the world wide web: a survey,”
ACM SIGMOD Record 27(3), 1998.
[7] Chunying Zhou,, Yaowen Zheng, “Query Rewriting from SPARQL to SQL for Relational Database
Integration” , IEIT Journal of Adaptive Dynamic Computing, Vol. 1, No. 1, 1-8
[8] Ora Lassila, and Ralph R. Swick, “Resource Description Framework (RDF) Model and Syntax
Specification” W3C Recommendation, 1999
[9] Dan Brickley, and R.V. Guha, “Resource Description Framework (RDF) Schema Specification”,
W3C Recommendation, 1999.
[10] Peng Liu, Xiaoying Wang, Aihua Bao and Xiaoxuan Wang , “Ontology Automatic Constructing
Based on Relational Database” Grid and Cooperative Computing (GCC), IEEE 9th International
2010.
[11] Xu Zhou, Guoji Xu andLei Liu , “An Approach for Ontology Construction Based on Relational
Database” , International Journal of Research and Reviews in Artificial Intelligence Vol. 1, No. 1,
March 2011.
[12] Olivier Corby, Leila Kefi-Khelif, Hacène Cherfi, Fabien Gandon and Khaled Khelif , “Querying the
Semantic Web of Data using SPARQL, RDF and XML”
Authors
Ayesha Banu is a Post Graduate in M.Sc (CS) from Kakatiya University in 2002
and M.Tech (CSE) from JNTUH in 2009 and she is pursuing her PhD from
JNTUH in the Area of “Semantic Web Mining”. She worked as a Lecturer in
Masterji PG college from 2003 to 2006 and as Asst. Professor in Alluri Institute of
Management Sciences from 2007 to 2009 and presently working as Associate
Professor in Department of Informatics in same college since May 2010. She is a
Life Member of ISTE and APSMS. She has delivered Guest Lectures in ACE
Engg. Colege ,MGIT , Deccan college of Engg and GNITS. On Research issues in
data mining , clustering techniques.
Dr. Syeda Sameen Fatima obtained B.Tech., Electronics and Communication Engineering from JNTU in
1982 , M.Phil., Computer Methods, University of Hyderabad, India in 1983 , M.S., Computer Science,
University of Massachusetts, Amherst, USA in 1993 and Ph.D. in Computer Science and Engineering, from
Osmania University, India in 2004. Her Research interests Include Information Retrieval Systems, Data
Mining ,Artificial Intelligence, Machine Learning. She ha 25 years of teaching experience and is presently
working as Professor and Head , Dept of CSE , Univ. College of Engineering , OU , Hyderabad. She has
published papers in various National and International Journals and Conferences.
Khaleel Ur Rahman Khan obtained B.E. (CSE) from Osmania University in 1993
and M.Tech (CS) from JNTU in 1998. PhD in Computer Science from Osmania
University in the area of Wireless Mobile Ad Hoc Networks. He is presently
working as Professor and Dean at ACE Engineering. His research interests include
Heterogeneous Networks, Opportunistic Networks, Transaction Management in Ad
Hoc and Sensor Networks, Data and Web Mining. He has published over 30 papers
in various Peer Reviewed International Journals and Conferences. He has teaching
experience of over 18 years.